The era of hand-rubbing videos has arrived
Author/ IT Times reporter Lin Fei
Editor/ Wang Xin, Sun Yan
Media practitioners have experienced the convenience and changes brought by AI models to the production process and creation.
Not long ago, I recommended an AI music creation website, app.suno.ai, to my colleague in charge of video production at the newspaper to solve the trouble of video soundtracks that had plagued her for a long time. In the past, if you purchased video background sound from the copyright library, the audio quality was often not high, and it did not match the video itself. However, if the BGM of the video editing platform is used, it will lead to the risk of infringement when distributing on multiple platforms.
app.suno.ai only need to enter the style, genre, content theme and other related Chinese keywords, and two two-minute songs can be automatically generated in just one minute, with lyrics and vocals, and the arrangement is also above the passing level. For short video production with moderate music quality requirements, AI-generated BGM undoubtedly provides a new option.
Personally, the AI assistance function I use most often is to run the interview recordings through the language model website, let the AI pre-sort them, and then manually proofread them, which directly shortens the interview recording sorting time by more than double. Next, I'm going to try using AI-generated images in the layout and manuscript images.
In recent years, in the field of content creation, consumer-grade artificial intelligence has shown a rapid development trend, and the "Wensheng Tu" large model represented by Stable Diffusion, Midjourney, DALL-E, etc., is reshaping the visual art creation pattern with unprecedented power. The widespread use of these technologies has disrupted the business model of commercial creative image galleries, and we can now access millions of AI-generated images under public domain licenses through sites such as Stockcake for free for a variety of projects, dramatically reducing the cost of acquiring high-quality visual assets.
At the same time, the field of music creation has also witnessed the rise of a series of AI products, such as MAGNet, Udio, NeuCo AI, and Soundful abroad, and NetEase Tianyin, Tencent TME Studio, BGM Cat, etc., all of which are committed to greatly reducing the threshold for music creation with the help of AI technology, so that non-professional users can easily create high-quality music works to meet the needs of various application scenarios.
Recently, Adobe announced that it will bring OpenAI-integrated Sora AI technology to its flagship Premiere Pro video editor, and while it may seem like a regular upgrade to video editing tools, it also means that AI technology has the power to revolutionize video creation and even "kill" traditional production models.
Judging from the official demo videos and user videos released by Sora, whether it is a person, or an animal; The real scene, or the CG mode, has fully achieved the effect of our common film and television dramas and commercials. To get this result, users only need to accurately describe the picture and scene they want, which lowers the professional threshold for video production.
At present, there is still a time limit for videos generated by ordinary users using Sora, mainly due to the constraints of the current large model in terms of storage and computing power. With the reduction of the computing cost of large models and the increasing evolution of AI tools, especially in the field of image and music generation, there is no need to rely on professional teams such as directors, actors, photographers, and lighting engineers in the future.
In the words of one netizen, it is "the arrival of the era of hand-rubbing movies".
AI content creation is not limited to mass production, but can also be highly customized. Users can provide detailed or abstract creative guidance to AI based on personal preferences, emotional states, specific situations, and other factors, so as to obtain unique works. In addition, AI can interact with users in real-time, dynamically adjusting the creative process, such as revising the plot of a script, adjusting the music style, or optimizing video clips in real time based on user feedback. This interactive authoring mode will greatly enhance user engagement. AI technology can not only directly generate movies based on text, but may also have the ability to imitate specific directorial styles, as if inviting masters such as Zhang Yimou, Jiang Wen, Hayao Miyazaki, and Quentin Tarantino to personally write "essays on the same topic".
Large models such as GPT-4 are still evolving. In the future, with the continuous optimization of technologies such as deep learning, reinforcement learning, and generative adversarial networks (GANs), AI will become more intelligent and flexible, and can respond to more diverse creative needs.
Debates about whether AI-generated content is "art" are all over the internet, and in my opinion, such discussions are redundant at a time when AI technology has had a disruptive impact on the form of artistic creation. As the OpenAI co-founder predicted in his TED Talk, the digital brain of the future is expected to be as good as the biological brain, and even surpass it in some ways. With the efficient processing of massive amounts of data, AI will surpass humans in terms of skills, creating works of art and ideas that surpass our existing imaginations.
In the face of the great changes caused by AI technology, it is difficult for both researchers and consumers to accurately foresee the future development path. What is certain, however, is that the content creation industry will usher in a deep transformation, and many jobs will be at risk of being replaced by AI.
In the face of this challenge, we should not passively avoid or maintain illusions, but should actively adapt to changes and continuously improve our innovative thinking and cross-border collaboration capabilities in an AI-assisted environment, so that we can not only enjoy the convenience and efficiency brought by technology, but also properly respond to potential challenges.
Typesetting / Ji Jiaying
Image / Adobe Sora
Source/"IT Times" official account vittimes
E N D
Please add a "star" so you don't miss us